WeAreDevelopers Logo WeAreDevelopers
Sign in
For You LIVE
WeAreDevelopers LIVE • Jun 17, 2020

The pitfalls of Deep Learning - When Neural Networks are not the solution

Adrian Spataru , Bohdan Andrusyak

Deep learning isn't a silver bullet for enterprise data. Discover why tree-based models often outperform neural networks and when to avoid their massive implementation costs.

Pause
Mute Enter Fullscreen
#1 about 3 min

Contrasting classical machine learning with deep learning approaches

An overview of the differences between manual feature extraction in classical algorithms and automatic extraction in neural networks.

#2 about 4 min

Successful applications and representation learning in neural networks

How deep learning drives breakthroughs in autonomous driving, language translation, and music generation through powerful feature representations.

#3 about 3 min

Real-world failures and unrealistic expectations of deep learning

How out-of-context translations and flawed facial recognition demonstrate the severe consequences of algorithmic limitations.

#4 about 4 min

Evaluating data quantity constraints in complex neural networks

Why deep learning struggles to identify accurate patterns when provided with limited or scarce data sets.

#5 about 2 min

The critical requirement for high quality training data

How garbage or incomplete data negatively impacts model accuracy regardless of algorithmic sophistication.

#6 about 4 min

Overcoming deep learning challenges with tabular business data

Why tree-based boosting methods consistently outperform neural networks on limited historical business spreadsheets.

#7 about 4 min

Comparing model explainability between algorithms and neural networks

The importance of understanding algorithm decisions and discarding irrelevant features by leveraging simple decision trees.

#8 about 2 min

Production complexity and engineering costs of oversized models

How excessive model constraints and large file sizes prevent improved architectures from being deployed into production.

#9 about 3 min

Analyzing computational resources and cloud training costs

The astronomical financial requirements needed to train massive natural language processing models from scratch.

#10 about 2 min

Calculating the business value and return on investment

Determining if marginal accuracy improvements from deep learning justify the additional technical complexity and overhead.

#11 about 2 min

Leveraging transfer learning to reduce data dependency issues

Reusing pre-trained models on new datasets accelerates development while avoiding high data and processing requirements.

#12 about 3 min

Exploring interpretability and future advancements in tabular processing

Discovering recent research around transparent tabular transformers and assessing readiness before integrating neural architectures.

Matching moments

1:35 min

Root causes of underlying AI initiative failures

Deivids Vilkinsons Deivids Vilkinsons +3 · World Congress 2026 Europe

1:58 min

Understanding core ingredients for successful AI adoption

Data Dan Data Dan · World Congress 2025

15:08 min

Audience questions on practical machine learning operational strategies

Lina Weichbrodt · LIVE

4:55 min

Business motivations and adoption challenges for generative AI

Maxim Salnikov Maxim Salnikov · LIVE

1:20 min

Uncovering the hidden technical debt in machine learning

Nico Axtmann · World Congress 2022

2:59 min

Introduction to AI experiments and production challenges

Amjad Shaikh Amjad Shaikh +1 · World Congress 2026 North America

More from WeAreDevelopers LIVE

Fake or News - Claude's DNA Discovery, Quick Cloning & Crime on Insta thumbnail
6:07
6 days ago

Fake or News - Claude's DNA Discovery, Quick Cloning & Crime on Insta

Andrew MacLean, Daniel Cranney, Christian Heilmann

WeAreDevelopers LIVE - Playing in Production thumbnail
1:06:40
6 days ago

WeAreDevelopers LIVE - Playing in Production

Andrew MacLean, Daniel Cranney, Christian Heilmann

WeAreDevelopers LIVE - Is CSS Catching Up to Flash? thumbnail
1:03:33
20 days ago

WeAreDevelopers LIVE - Is CSS Catching Up to Flash?

Chris Heilmann, Daniel Cranney, Bramus Van Damme

WeAreDevelopers LIVE - Project Chopin: Make Your Team and Agents Play Along thumbnail
1:21:17
27 days ago

WeAreDevelopers LIVE - Project Chopin: Make Your Team and Agents Play Along

Chris Heilmann, Daniel Cranney, Krzysztof Cieślak

Fake or News - AI Toothbrushes, Carrier Pigeon Tech & Life on Mars - Krzysztof Cieślak thumbnail
4:23
29 days ago

Fake or News - AI Toothbrushes, Carrier Pigeon Tech & Life on Mars - Krzysztof Cieślak

Chris Heilmann, Daniel Cranney

WeAreDevelopers LIVE - Open Source Agent Toolkits thumbnail
1:00:38
about 1 month ago

WeAreDevelopers LIVE - Open Source Agent Toolkits

Chris Heilmann, Daniel Cranney, Eddie Jaoude

Related videos

Big Business, Big Barriers? Stress-Testing AI Initiatives. thumbnail
27:53
World Congress 2024 • about 2 years ago

Big Business, Big Barriers? Stress-Testing AI Initiatives.

Marin Niehues

Getting Started with Machine Learning thumbnail
44:11
LIVE • about 5 years ago

Getting Started with Machine Learning

Alexandra Waldherr

The Limits of LLMs in Real-World Applications thumbnail
31:07
World Congress 2026 Europe • 3 months ago

The Limits of LLMs in Real-World Applications

Deivids Vilkinsons, Guillaume Esnou, Laura Moritz, Mariam Hakobyan

AI beyond the code: Master your organisational AI implementation. thumbnail
30:49
LIVE • almost 2 years ago

AI beyond the code: Master your organisational AI implementation.

Marin Niehues

Fireside Chat: Deep Learning, Deep Impact: Harnessing AI for Language Innovation thumbnail
31:15
World Congress 2023 • about 3 years ago

Fireside Chat: Deep Learning, Deep Impact: Harnessing AI for Language Innovation

Jaroslaw Kutylowski, Rudi Bauer

AI That Fits Your Business, Not the Other Way Around thumbnail
25:56
World Congress 2026 Europe • 3 months ago

AI That Fits Your Business, Not the Other Way Around

Anshul Jindal, Cansu Kavili Örnek

Related articles

From learning to earning

Jobs that call for the skills explored in this talk.